Comparison of Different Classifier Performances for Condition Monitoring of Induction Motor Using DWT

Abstract

Condition monitoring of induction motors has been one of the prime focused area in recent years for its utility and effectiveness. The main aim of this article is to find out induction motor faults in incipient stage, so that uninterrupted hazard free operation could be achieved. It is found that different types of faults could not be precisely identified by Motor Current Signature Analysis (MCSA) using FFT analysis alone. In this proposed work, fault related information is extracted from Park’s Vector Modulus (PVM) current using Discrete Wavelet Transform (DWT). Different DWT coefficients are generated with specific time-frequency resolution to segregate different fault information features. DWT coefficient helps to generate more fault features using statistical methods. Overall, this method shows reasonable accuracy in fault identification by implementing different types of classifiers.